
import numpy as np 
import pandas as pd 
import matplotlib.pyplot as plt
from sklearn.linear_model import LinearRegression
import seaborn as sns
import plotly.express as px
import matplotlib.pyplot as plt


df=pd.read_csv('KosteckiDillon.csv')


df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 4152 entries, 0 to 4151
Data columns (total 10 columns):
 #   Column      Non-Null Count  Dtype  
---  ------      --------------  -----  
 0   Unnamed: 0  4152 non-null   int64  
 1   id          4152 non-null   int64  
 2   time        4152 non-null   int64  
 3   dos         4152 non-null   int64  
 4   hatype      4152 non-null   object 
 5   age         4152 non-null   int64  
 6   airq        4152 non-null   float64
 7   medication  4152 non-null   object 
 8   headache    4152 non-null   object 
 9   sex         4152 non-null   object 
dtypes: float64(1), int64(5), object(4)
memory usage: 324.5+ KB


df.drop(["Unnamed: 0"], axis=1, inplace=True)


df['medication'].value_counts()

continuing    2386
reduced        981
none           785
Name: medication, dtype: int64


fig = px.pie(df, values='medication', names='age',title='pie chart between age and medication')
fig.show()


fig = px.bar(df, x='age', y='medication',color='age',title='graph between age and medication')
fig.show()


df_headache_yes = df[df['headache'] == 'yes']['age']
df_headache_no = df[df['headache'] == 'no']['age']


print(df_headache_yes)

0       30
1       30
2       30
3       30
4       30
        ..
4143    26
4144    26
4146    26
4147    26
4151    26
Name: age, Length: 2666, dtype: int64


print(df_headache_no)

12      30
22      36
47      36
66      28
68      28
        ..
4108    43
4145    26
4148    26
4149    26
4150    26
Name: age, Length: 1486, dtype: int64


df.hatype.unique()

array(['Aura', 'No Aura', 'Mixed'], dtype=object)


df.medication.unique()

array(['continuing', 'reduced', 'none'], dtype=object)


df.headache.unique()

array(['yes', 'no'], dtype=object)


df.sex.unique()

array(['female', 'male'], dtype=object)


df[df.isnull().any(axis=1)]


sns.pairplot(df[['id', 'time', 'dos', 'hatype', 'age', 'airq', 'medication', 'headache','sex']])

<seaborn.axisgrid.PairGrid at 0x1dc957bee50>


df.corr()


df.columns

Index(['id', 'time', 'dos', 'hatype', 'age', 'airq', 'medication', 'headache',
       'sex'],
      dtype='object')


df.dtypes

id              int64
time            int64
dos             int64
hatype         object
age             int64
airq          float64
medication     object
headache       object
sex            object
dtype: object


sex=df.sex.replace(to_replace=['female', 'male'], value=[0,1])
sex

0       0
1       0
2       0
3       0
4       0
       ..
4147    1
4148    1
4149    1
4150    1
4151    1
Name: sex, Length: 4152, dtype: int64


df['sex']=pd.Series(sex)


df.dtypes

id              int64
time            int64
dos             int64
hatype         object
age             int64
airq          float64
medication     object
headache       object
sex             int64
dtype: object


df.headache.unique()

array(['yes', 'no'], dtype=object)


headache=df.headache.replace(to_replace=['no', 'yes'], value=[0,1])


df['headache']=pd.Series(headache)

df


df.medication.unique()

array(['continuing', 'reduced', 'none'], dtype=object)


medication=df.medication.replace(to_replace=['none','reduced','continuing'],value=[0,1,2])


df['medication']=pd.Series(medication)

df


len(df.hatype.unique())

3


hatype=df.hatype.replace(to_replace=['Aura','No Aura','Mixed'],value=[1,0,2])


df['hatype']=pd.Series(hatype)


df_final=df
df_final


df_final.corr()


df_final['medication'].value_counts()

2    2386
1     981
0     785
Name: medication, dtype: int64


data_majority = df_final[df_final.medication==2]
data_minority2 = df_final[df_final.medication==1]
from sklearn.utils import resample
data_minority_upsampled1 = resample(data_minority2, 
                                 replace=True,     # sample with replacement
                                 n_samples=2000,    # to match majority class
                                 random_state=4) # reproducible results
 
# Combine majority class with upsampled minority class
data_upsampled = pd.concat([data_majority, data_minority_upsampled1])
 
# Display new class counts
#data_upsampled.balance.value_counts()
data_upsampled['medication'].value_counts()

2    2386
1    2000
Name: medication, dtype: int64


data_majority = df_final[df_final.medication==2]
data_minority2 = df_final[df_final.medication==0]
from sklearn.utils import resample
data_minority_upsampled1 = resample(data_minority2, 
                                 replace=True,     # sample with replacement
                                 n_samples=2000,    # to match majority class
                                 random_state=4) # reproducible results
 
# Combine majority class with upsampled minority class
data_upsampled = pd.concat([data_upsampled, data_minority_upsampled1])
 
# Display new class counts
#data_upsampled.balance.value_counts()
data=data_upsampled['medication'].value_counts()


data=data_upsampled.copy()


corr = df_final.corr()
plt.figure(figsize=(10,8))
sns.heatmap(corr,annot=True)

<AxesSubplot:>


from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.linear_model import LinearRegression
from sklearn.neighbors import KNeighborsRegressor
from sklearn.ensemble import RandomForestRegressor
from sklearn.ensemble import AdaBoostClassifier
from sklearn.svm import SVC
from sklearn.metrics import classification_report, confusion_matrix,accuracy_score
from sklearn.preprocessing import StandardScaler 
from sklearn.model_selection import cross_val_score


X=data.drop(['medication'],1)
y=data['medication']


df_final


X_train, X_test, y_train, y_test = train_test_split(X,y, test_size= 0.3, random_state=4)


X2 = data.drop(['medication'], axis=1)
y2 = data['medication']


X_train2, X_test2, y_train2, y_test2 = train_test_split(X2, y2, test_size=0.2,random_state=4)


scaler = StandardScaler()
scaler.fit(X_train2)

X_train2 = scaler.transform(X_train2)
X_test2 = scaler.transform(X_test2)


abc = AdaBoostClassifier(n_estimators=50,learning_rate=1)
model = abc.fit(X_train2, y_train2)
y_pred2 = model.predict(X_test2)


print("Accuracy:",accuracy_score(y_test2, y_pred2))

Accuracy: 0.7355242566510172


from sklearn.model_selection import cross_val_score  
print(cross_val_score(abc,X2,y2,cv=5,scoring="accuracy").mean())

0.43987705927551735


X3 = data.drop(['medication'], axis=1)
y3 = data['medication']


from sklearn.model_selection import train_test_split
X_train3, X_test3, y_train3, y_test3 = train_test_split(X3,y3,test_size=0.2,random_state=4)


scaler = StandardScaler()
scaler.fit(X_train3)

X_train3 = scaler.transform(X_train3)
X_test3 = scaler.transform(X_test3)


clf = KNeighborsRegressor()   
clf.fit(X_train3, y_train3) 
y_pred3 = clf.predict(X_test3)
print(clf.score(X_test3, y_test3))

0.8318733186406381


#print(cross_val_score(clf,X3,y3,cv=4,scoring="accuracy").mean())


#from sklearn import metrics
#print(metrics.classification_report(y_test3, y_pred3))


X1 = data.drop(['medication'], axis=1)
y1 = data['medication']


from sklearn.model_selection import train_test_split
X_train1, X_test1, y_train1, y_test1 = train_test_split(X1,y1,test_size=0.2,random_state=4)


scaler = StandardScaler()
scaler.fit(X_train1)

X_train1 = scaler.transform(X_train1)
X_test1 = scaler.transform(X_test1)


from sklearn.ensemble import RandomForestClassifier 
clfe = RandomForestClassifier(n_estimators = 50)   
clfe.fit(X_train1, y_train1) 
y_pred1 = clfe.predict(X_test1)
print(clfe.score(X_test1, y_test1))

1.0


from sklearn import metrics
from sklearn.metrics import confusion_matrix
cmx = metrics.confusion_matrix(y_test1,y_pred1)
cmx

array([[419,   0,   0],
       [  0, 377,   0],
       [  0,   0, 482]], dtype=int64)


print(cross_val_score(clfe,X1,y1,cv=4,scoring="accuracy").mean())

0.668188846411583


from sklearn import metrics
print(metrics.classification_report(y_test1, y_pred1))

              precision    recall  f1-score   support

           0       1.00      1.00      1.00       419
           1       1.00      1.00      1.00       377
           2       1.00      1.00      1.00       482

    accuracy                           1.00      1278
   macro avg       1.00      1.00      1.00      1278
weighted avg       1.00      1.00      1.00      1278


param_grid = {'C': [0.1, 1, 10, 100, 1000], 
              'kernel': ['rbf']}
grid = GridSearchCV(SVC(), param_grid, refit = True, verbose = 3) 
  
# fitting the model for grid search 
grid.fit(X_train, y_train)

Fitting 5 folds for each of 5 candidates, totalling 25 fits
[CV] C=0.1, kernel=rbf ...............................................

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.5s remaining:    0.0s

[CV] ................... C=0.1, kernel=rbf, score=0.431, total=   0.6s
[CV] C=0.1, kernel=rbf ...............................................
[CV] ................... C=0.1, kernel=rbf, score=0.427, total=   0.6s
[CV] C=0.1, kernel=rbf ...............................................

[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    1.0s remaining:    0.0s

[CV] ................... C=0.1, kernel=rbf, score=0.453, total=   0.8s
[CV] C=0.1, kernel=rbf ...............................................
[CV] ................... C=0.1, kernel=rbf, score=0.436, total=   0.9s
[CV] C=0.1, kernel=rbf ...............................................
[CV] ................... C=0.1, kernel=rbf, score=0.459, total=   0.8s
[CV] C=1, kernel=rbf .................................................
[CV] ..................... C=1, kernel=rbf, score=0.477, total=   0.5s
[CV] C=1, kernel=rbf .................................................
[CV] ..................... C=1, kernel=rbf, score=0.453, total=   0.5s
[CV] C=1, kernel=rbf .................................................
[CV] ..................... C=1, kernel=rbf, score=0.446, total=   0.6s
[CV] C=1, kernel=rbf .................................................
[CV] ..................... C=1, kernel=rbf, score=0.469, total=   0.6s
[CV] C=1, kernel=rbf .................................................
[CV] ..................... C=1, kernel=rbf, score=0.487, total=   0.6s
[CV] C=10, kernel=rbf ................................................
[CV] .................... C=10, kernel=rbf, score=0.534, total=   0.6s
[CV] C=10, kernel=rbf ................................................
[CV] .................... C=10, kernel=rbf, score=0.506, total=   0.8s
[CV] C=10, kernel=rbf ................................................
[CV] .................... C=10, kernel=rbf, score=0.477, total=   0.8s
[CV] C=10, kernel=rbf ................................................
[CV] .................... C=10, kernel=rbf, score=0.530, total=   0.8s
[CV] C=10, kernel=rbf ................................................
[CV] .................... C=10, kernel=rbf, score=0.501, total=   0.5s
[CV] C=100, kernel=rbf ...............................................
[CV] ................... C=100, kernel=rbf, score=0.577, total=   0.7s
[CV] C=100, kernel=rbf ...............................................
[CV] ................... C=100, kernel=rbf, score=0.550, total=   0.6s
[CV] C=100, kernel=rbf ...............................................
[CV] ................... C=100, kernel=rbf, score=0.527, total=   0.8s
[CV] C=100, kernel=rbf ...............................................
[CV] ................... C=100, kernel=rbf, score=0.596, total=   1.0s
[CV] C=100, kernel=rbf ...............................................
[CV] ................... C=100, kernel=rbf, score=0.541, total=   1.0s
[CV] C=1000, kernel=rbf ..............................................
[CV] .................. C=1000, kernel=rbf, score=0.638, total=   1.1s
[CV] C=1000, kernel=rbf ..............................................
[CV] .................. C=1000, kernel=rbf, score=0.636, total=   1.2s
[CV] C=1000, kernel=rbf ..............................................
[CV] .................. C=1000, kernel=rbf, score=0.647, total=   1.6s
[CV] C=1000, kernel=rbf ..............................................
[CV] .................. C=1000, kernel=rbf, score=0.655, total=   1.2s
[CV] C=1000, kernel=rbf ..............................................
[CV] .................. C=1000, kernel=rbf, score=0.642, total=   1.3s

[Parallel(n_jobs=1)]: Done  25 out of  25 | elapsed:   20.0s finished

GridSearchCV(estimator=SVC(),
             param_grid={'C': [0.1, 1, 10, 100, 1000], 'kernel': ['rbf']},
             verbose=3)


# print best parameter after tuning 
print(grid.best_params_) 

# print how our model looks after hyper-parameter tuning 
print(grid.best_estimator_)

{'C': 1000, 'kernel': 'rbf'}
SVC(C=1000)


grid_predictions = grid.predict(X_test) 

# print classification report 
print(classification_report(y_test, grid_predictions))

              precision    recall  f1-score   support

           0       0.71      0.65      0.68       623
           1       0.67      0.63      0.65       569
           2       0.58      0.65      0.61       724

    accuracy                           0.65      1916
   macro avg       0.66      0.64      0.65      1916
weighted avg       0.65      0.65      0.65      1916

	id	time	dos	age	airq
id	1.000000	-0.137080	0.971239	0.057145	-0.102202
time	-0.137080	1.000000	-0.085977	-0.055781	-0.002529
dos	0.971239	-0.085977	1.000000	0.021148	-0.129934
age	0.057145	-0.055781	0.021148	1.000000	-0.009163
airq	-0.102202	-0.002529	-0.129934	-0.009163	1.000000

	id	time	dos	hatype	age	airq	medication	headache	sex
id	1.000000	-0.137080	0.971239	0.107066	0.057145	-0.102202	0.000987	-0.019063	0.176734
time	-0.137080	1.000000	-0.085977	0.122126	-0.055781	-0.002529	-0.190634	-0.038022	0.169886
dos	0.971239	-0.085977	1.000000	0.145901	0.021148	-0.129934	0.013337	-0.005127	0.168754
hatype	0.107066	0.122126	0.145901	1.000000	-0.151747	-0.120952	-0.056966	0.124049	0.066247
age	0.057145	-0.055781	0.021148	-0.151747	1.000000	-0.009163	-0.039558	-0.058954	0.085890
airq	-0.102202	-0.002529	-0.129934	-0.120952	-0.009163	1.000000	-0.050362	0.014357	-0.015205
medication	0.000987	-0.190634	0.013337	-0.056966	-0.039558	-0.050362	1.000000	0.120437	-0.171292
headache	-0.019063	-0.038022	-0.005127	0.124049	-0.058954	0.014357	0.120437	1.000000	-0.003917
sex	0.176734	0.169886	0.168754	0.066247	0.085890	-0.015205	-0.171292	-0.003917	1.000000

Importing modules¶

confussion matrix¶

adaboost¶

KNeighbour¶

Randomforest¶

svm¶

	id	time	dos	hatype	age	airq	medication	headache	sex
0	1	-11	753	Aura	30	9.0	continuing	1	0
1	1	-10	754	Aura	30	7.0	continuing	1	0
2	1	-9	755	Aura	30	10.0	continuing	1	0
3	1	-8	756	Aura	30	13.0	continuing	1	0
4	1	-7	757	Aura	30	18.0	continuing	1	0
...	...	...	...	...	...	...	...	...	...
4147	133	44	1232	Aura	26	20.0	continuing	1	1
4148	133	45	1233	Aura	26	27.0	continuing	0	1
4149	133	46	1234	Aura	26	28.0	continuing	0	1
4150	133	47	1235	Aura	26	33.0	continuing	0	1
4151	133	48	1236	Aura	26	25.0	continuing	1	1